I still hear people saying things like, “AI can never write as well as a human copywriter.”
Or, “I asked ChatGPT to write an article, and it was terrible.”
Respectfully, these people are missing the point.
Just because AI can’t entirely replace the full set of skills of a human writer doesn’t mean it can’t be valuable.
Look at this through a different lens…
People get excited by how AI models like ChatGPT make it easy to create copy and content at scale.
As a writer you can create way more content when using AI. And yes, it’s a lot easier. No more writer’s block. AI can help you come up with new ideas for content, suggest headlines, and even write a finished draft for you.
And instead of a new post or article taking you two or three hours, or more, you can get it done in 5 to 10 minutes. How awesome is that?
Well, it IS awesome.
But…
Writing prompts shouldn’t feel like you’re acting as a “prompt engineer”.
Honestly, I don’t much like that term.
First, I think it’s disrespectful to true engineers. Writing prompts is not engineering.
But more to the point, I think the title steers us in the wrong direction. Engineering implies a fixed process and a certain formality. A process with a guaranteed and constant outcome. “Use these top 20 prompts for optimum results.”
That approach sounds attractive, but I don’t think it’s the way to get the best outputs from an AI model like ChatGPT.
Much of the time, we use tools like ChatGPT to answer questions.
Search engines like Google have trained us for decades to interact with the internet in this question-and-answer format.
“Where can I find the best price for an all-inclusive vacation in Cancun?”
“How do I make coffee with a percolator?”
We bring those same habits to models like ChatGPT, or hybrid tools like Perplexity.ai.
You have a lot of choices when using AI to write copy and content.
First, you can choose between the major Large Language Models, like ChatGPT, Gemini, Claude3 and others.
Then you have dozens, if not hundreds of new AI-writing platforms, most of which are using these models as the engines that drives them.
It’s little wonder the web is being flooded with new AI content, most of it sounding like it was written by a well-meaning robot.
The content isn’t badly written. With some decent prompts, most of these tools will give you a very passable first and second draft.
But there IS a problem.
This is what Sam Altman, CEO of OpenAI, said in an interview last fall:
“It will mean that 95% of what marketers use agencies, strategists, and creative professionals for today will easily, nearly instantly and at almost no cost be handled by the AI — and the AI will likely be able to test the creative against real or synthetic customer focus groups for predicting results and optimizing. Again, all free, instant, and nearly perfect. Images, videos, campaign ideas? No problem.”
To put this into context, he’s not talking about GPT-4 or the upcoming GPT-5. He’s talking about when AI achieves Artificial General Intelligence, AGI.
AGI is broadly defined as the stage at which an AI model gains enough skills to perform any task that humans can do, with equal or better proficiency.
We’re not there yet.